Gesture-Based Recognition of ISL to Inclusive Communication in Marathi Text

Abstract

Hand gestures are the media of individuals with total hearing loss but not useful to connect with ordinary person. The hand gesture language recognition becomes most potential area of research. In the present research work the MSL recognition system is developed with help of mediaPipe for real time recognition. It is not possible every time to get the interpreter for translation and understanding of Marathi sign language for common people. In the previous research work many researchers had undertakes the machine learning for classification and extraction of feature in the light of development of said system. They have faced several challenges in these research areas. The present research work will overcome the challenges of these research areas. As a result, it is found that by using direct image data and by using Euclidean distance function the ANN model shows the higher accuracy (approx. 95. %) than logistics regression, decision tree and SGD Model.